UCRL-53719

Radiation Chemical Effectsin Experiments to Study

the Reaction of Glassin an Environment ofGamma-Irradiated Air,Groundwater, and Tuff

R. A. Van Konynenburg

May 2, 1986

CPA

i•T/ L - 7-

DISCLAIMER

This document was prepared as an account of work sponsored by an agency of the United States GovernmenLNeither the United States Government nor the University of California nor any of their employees, makes anywarranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness orusefulness of any information, apparatus, product, or process disclosed, or represents that its use would not nfringeprivately owned rights. Reference herein to any specific commercial products, process, or servie by trade nametrademark, manufacturer, or otherwise, does not necessariy constitute or imply its endorsement, recommendation. orfavoring by the United States Government or the University of California. The views and opinions of authorsexpressed herein do not necessarily state or reflect those of the United States Government or the University ofCalifornia, and shall not he used for advertising or product endorsement purposes.

Prepared by Nevada Nuclear Waste Storage Investigations (NWSI) Projectparticipants as part of the Civilian Radioactive Waste Management Program.The NNWSI Project is managed by the Waste Management Project Office of theU.S. Department of Energy, Nevada Operations Office. NNWSI Project work issponsored by the Office of Geologic Repositories of the DOE Office of CivilianRadioactive Waste Management. 1

Work performed under the auspices of the US. Department of Energy by Lawrence Livermore National Laboratoryunder Contract W-7405-En48.

UCRL-53719Distribution Category UC-70

Radiation Chemical Effectsin Experiments to Study

the Reaction of Glassin an Environment of

Gamma-Irradiated Air,Groundwater, and Tuff

R. A. Van Konynenburg

l Manuscript date: May 2, 1986

LAWRENCE LIVERMORE NATIONAL LABORATORY IlLUniversity of California Livermore, California * 94550 W

Available from: National Technical Information Service * U.S. Department of Commerce5285 Port Royal Road * Springfield, VA 22161 * A03 * (Microfiche AO)

TABLE OF CONTENTS

1. Introduction...................................... ...................................... 3

2. Description of Experiments and Results................................ 5

3. Drisc sso ................................................................... 15

3.2 Changes in Composition due to Heating........................... 16

3.3 Overview of Major Features and Expected Observable Results ...... 17

3.4 Gas Pae................................... 19

3 .5 L iqu id Phase ................................... 2 0

3.6 Gas and Liquid Phases Considered Together....................... 29

3.7 Interactions with the Solid Phase .................... ........... 30

3.8 Decomposition of Water.......................................... 32

3.9 Fixation of Nitrogen.. ....... ................................... 33

3.10 Hydrogen Ion Production ......................................... 36

3.11 Nitrite-Nitrate Ratio ............................. ' 36

3.12 Likely Effects on Glass Reaction .............................. 40

3.13 High Gas-to-Liquid Volume Ratio Experiments ..... ........... .... 40

3. 14 Anomalo us Tes t s itoy. .. . . . . . . . . . . . . . . . . . . . . . . . . . . .... 42

4. Conclusions .............................6.................... 6 ..................................... 45

5. Application to the Repository ................................................. 47

R e f e r e n c e s . * ... * . ... . .................... . ........................... 5 1

iii

iv

RADIATION CHEMICAL EFFECTS IN EXPERIMENTS TO STUDY

THE REACTION OF GLASS IN AN ENVIRONMENT OF

GAMMA-IRRADIATED AIR, GROUNDWATER, AND TUFF

ABSTRACT

The results of experiments performed by John K. Bates et al. on the

reaction of nuclear waste glass with a gamma-irradiated 900C aqueous solution

were analyzed using theory developed from past research in radiation

chemistry. The aqueous solution they used is similar to what would be

expected in a water-saturated environment in a nuclear waste repository in

tuff. The purpose of our study was to develop an understanding of the

radiation-chemical processes that occurred in the Bates et al. experiments so

the results could be applied to the design and performance analysis of a

proposed repository in unsaturated tuff in Nevada. For the Bates et al.

experiments at the highest dose (269 Mrad), which originally coltained about

16 ml of equilibrated" water taken from Nevada Test Site Well J-13 and 5.4 ml

of air, we predicted that water decomposition to H2 and 02 would produce a

pressure increase of at least 1.0 MPa at 200C. We also predicted that nitrogen

fixation from the air would occur, producing an increase of 1.6 x 10-4 M in

total fixed nitrogen concentration in solution. In addition, an equimolar

production of Hi would occur, which would be buffered by the OO in the3water. The fixed nitrogen in solution was predicted to be present as NO- and

NO with the ratio influenced by the presence of materials catalytic to the

decomposition of H202. We found reasonable agreement between our predictions

and the observations of Bates et al., where comparisons were possible. We

apply the results to the proposed Nevada repository to the degree possible,

given the different expected conditions.

1

2

1. INTRODUCTION

The Nevada Nuclear Waste Storage Investigations (NNWSI) Project is

evaluating the tuff deposits at Yucca Mountain in Nye County, Nevada, as a

possible location for a high-level nuclear waste repository. Lawrence

Livermore National Laboratory (LLNL) is responsible for waste package

development for this site. An important aspect of this development is to

predict the extent of release into the environment of radionuclides from the

nuclear waste expected at the site. The waste form planned for the first

defense high-level nuclear waste to be emplaced (as well as for the waste from

the West Valley, New York, facility) is borosilicate glass. Accordingly, we

are studying the behavior of borosilicate glass under conditions similar to

those in the expected tuff repository environment.

We anticipate that the repository will be located in the unsaturated zone

above the water table. We also expect that during the first 300 to 1000 years

the waste form will be protected by sealed outer containers, and that the

thermal power density from nuclear decay will be sufficient to prevent liquid

water from coming into contact with most of the packages during this period.

The maximum gamma dose rate near waste packages containing the glass waste

form is expected to be about 5.5 x 103 rem/hour at the time of emplacement.1

The dominant gamma emitter during the containment period will be 137Cs, with a

half-life of 30.17 years. The gamma dose rate would thus decay by about 10

half-lives during a 300-year period, which is equivalent to about three orders

of magnitude. The resulting dose rate would be a few rem/hour maximum after

300 years. Nevertheless, there is a small probability that some liquid water

could contact some of the glass during the time when a significant gamma

radiation dose rate was still present.

Past studies have shown that gamma irradiation can increase the rate and

extent of reactions of nuclear waste glasses with deionized water and aqueous

solutions compared to what is observed in the absence of radiation; Pederson

and McVay provide a recent review of these studies.2 The largest increases

have been seen when air was in contact with the solution, because of the

production of nitric acid during irradiation and the resulting lowering of the

pH. 3-6 Smaller increases have been attributed to the transient radical

species produced by irradiation of the water, particularly the hydroxyl

radical O. 7

3

We expect that a repository located in tuff at Yucca Mountain would be

open to gaseous exchange with the atmosphere because of the fractures and

connected porosity of the rock. Therefore the potential for nitric acid

production would exist. The presence of dissolved C02 also raises the

possibility of production of carboxylic acids, which can act as chelating

agents increasing corrosion and dissolution rates.8-12 However, the expected

water composition includes bicarbonate as the major anion, which would tend to

buffer any tendency toward a more acid pH. In addition, where water is in

contact with tuff, considerable buffering would be supplied by the feldspar

minerals, which readily exchange Na+, K+, and Ca++ for HI ions.13-17 The

simultaneous presence of oxygen and neutral-to-alkaline pH could also reduce

the production of carboxylic acids (see Section 3.5).

Because of the existence of these opposing factors and the lack of a

complete theoretical treatment of the glass-water interaction process,

particularly under irradiation, John K. Bates et al. of Argonne National

Laboratory designed and carried out appropriate laboratory experiments to

study the glass behavior under saturated tuff conditions with gamma

irradiation. Some of the results have been reported.18'19

In order to apply these results to the prediction of glass behavior in a

repository, it is necessary to understand the detailed mechanisms of the

processes that occurred during the Bates et al. experiments. Therefore, using

theory developed in earlier studies, in this paper we predict the radiation

chemical effects that would be expected to occur in the experimental con-

figuration used by Bates et al. and compare these predictions with parameters

observed by them.

The purpose of this work is to contribute to the development of a detailed

understanding of the behavior of nuclear waste glass in a tuff repository.

The repository can then be designed and its performance projected in such a

way that the required reasonable assurance can be provided that radionuclide

releases will be below the limits established by the Environmental Protection

Agency2O and the Nuclear Regulatory Commission21 (40CFRl91 and lOCFR60).

4

2. DESCRIPTION OF EXPERIMENTS AND RESULTS

Bates et al. performed two sets of experiments to study the influence of

gamma irradiation on the reaction of actinide-doped Savannah River Laboratory

165-frit nuclear waste glass in a saturated tuff environment.18'19'22 These

experiments consisted of irradiating small vessels containing various

combinations of glass, tuff, water, and air with gamma rays for several time

periods while maintaining an elevated temperature. The dissolved solutes and

the glass surfaces were then analyzed.

The vessels were 304L stainless steel Parr reaction vessels, sealed with

silicone rubber gaskets. They had a volume of 21.4 cm3. The nominal

composition of 304L stainless steel is given in Table 1. The complete gasket

composition was not measured. Bates found only Si in an examination by

energy-dispersive x-ray analysis, which was sensitive only to elements of

Z 2!11.22 A later 1500C leaching experiment on the gasket material with HNO3

solution revealed only Si as a major constituent, with traces of B, Mg, and Ca

also released. The vessels were machined using Trimsol, a soluble oil-type

cutting fluid manufactured by the Master Chemical Company. They were then

polished using an aluminum oxide slurry. The vessels were ultrasonically

cleaned in acetone, rinsed three times in high purity water, submerged in 900C

1% HNO3 solution for one hour, rinsed again three times with high purity

water, and dried. The gaskets were heated to 900C in high purity water for

two hours, rinsed with high purity water, and dried.

The glass consisted of two types. The first was Savannah River

Laboratory (SRL) 165 black frit to which U, Cs, and Sr had been addedl it is

referred to as SRL U glass. The second was SRL U glass to which 237Np,239Pu, and 241Am had been added; it is referred to as SRL A glass. In the

first set of experiments, the glass was used both in the form of discs and

powder. Discs, when used, were supported on 304L stainless steel stands. In

the second set of experiments, only discs were used. The composition of the

glass used in the first set of tests is given in Table 2 the actinide

composition was slightly different in the second set.

The first set of experiments used crushed Topopah Spring tuff from a

drill hole at Fran Ridge near the Yucca Mountain site in Nevada (UE-25 h 1,

173.0 to 173.6). We expect the major and minor element whole rock composition

of this tuff to be similar to that reported by ielinski23 for drill hole

5

Table 1. Composition of AISI standard type 304L stainless steel.a

Element Weight %

Iron

Chromium

Nickel

Manganese

Silicon

Phosphorus

Sulfur

Carbon

Balance (at least 65)

18.0-20.0

8.0-12.0

2.0 maximum

1.0 maximum

0.045 maximum

0.03 maximum

0.03 maximum

a From American Society for Metals, Metals Handbook (A.S.M., Metals Park,

OH, 1980) 9th ed., vol.3, p. 5.

6

Table 2. Composition of SRL glasses used in first set of tests.a

Analysis (weight %)SRL 165 U/A Black Frit

Component ANLb FerroC MCCdA1203 4.08 4.1 4.3

B203 6.76 6.8 6.8

BaO 0.06 < 0.1

CaO 1.62 1.5 1.6

Fe2 03 11.35 12.3 11.7

FeO 0.35

K20 NA 0.2

Li2O 4.18 4.7 4.8

MgO 0.70 0.8 0.6

MnO 2.27 2.9 2.8

Na2O 10.85 10.3 10.8

NiO 0.85 0.9 0.8

P205 0.3

SiO2 52.86 54.1 51.6

TiO2 0.14 0.2

ZnO 0.04 <0.1

Zr02 0.66 1.2 0.7

F NA 0.06Cl NA 0.05Pb NA 0.05

2NP02 0.0082 3 9 PuO2 0.022e

2 4 1 Am203 0.00036e

U308 0.96

Cs2O 0.072

SrO 0.11

Othersf

NOTE: NA - not analyzed.aFrom J. K. Bates, D. F. Fischer, and T. J. Gerding, The Reaction of Glass

in a Gamma Irradiated Saturated Tuff Environment. Part 1: SRL 165 Glass,Argonne National Laboratory, Argonne, IL, ANL-85-62 (1985).

Only SRL 3 lass 2 As analyze 1 for all components. SRL A glass wasanalyzed for Np, Pu, and Am. The composition of both glasses areassumed to be the same, except for actinide elements.

c Black frit supplied to ANL by SRL, composition as determined by Ferro Corp.d Black frit supplied by SRL to the MCC, composition as determined by MCC.e Analysis by alpha spectrometry.f La203, Nd2O3, CeO2 < 0.05S M003 < 001 and Cr2 03 < 0.01 weight %.

7

USW-G1, given in Table 3. The major mineral species present in this rock are

quartz, cristobalite, and alkalai and plagioclase feldspars. The second set

of experiments used discs of tuff from USW-G1 (1235.1 - 5 ft.).

"Equilibrated -13 water" was so-called because it was prepared by

heating water from Well J-13 near Yucca Mountain, together with crushed tuff,

in a closed Teflon bottle for two weeks at 900C, filtering it, and storing it

in a closed polypropylene container. Its composition should therefore be

closer to what would exist in thermodynamic equilibrium with tuff at 900C than

the starting composition of J-13 water. The composition of the water used in

the first set of experiments is shown in Table 4. The bicarbonate and

potassium concentrations were not measured, so estimates were made based on

anion-cation balance and previous analyses of solutions prepared in a similar

way. The dissolved gas concentrations were calculated using Henry's Law, and

the carbonate concentration was calculated assuming equilibrium (constants are

given in Table 5). The water composition was slightly different for the

second set of experiments. Measured pH values for the starting water used in

both sets are shown in Table 6. The difference between the two measured

values probably results from nonequilibrium concentrations of dissolved CO2.

The air was ordinary laboratory air, the dry composition of which was

probably near to that of standard dry air, as shown in Table 7. The humidity

of the initial air was not controlled, but the air in the sealed capsules

would have reached 100% relative humidity because there was excess liquid

water present, and there was sufficient time to reach equilibrium vapor

pressure.

The gamma irradiation was provided by 6o sources. In the first set of

experiments, the dose rate (measured at the beginning of the experiments) was

2.0 + 0.2 x 105 rad/hour inside the vessels. In the second set of

experiments, the dose rate was 1.0 + 0.2 x 104 rad/hour. The vessels were

arranged relative to the sources to achieve essentially the same dose rates in

all the vessels. (Additional experiments at lower dose rates are underway.)

The vessels were placed inside an oven, the temperature of which was

controlled at 90.0 + 0.50C. However, post-test measurements revealed that the

actual vessel temperature was about 920C for the first set of experiments.

The vessels in the first set of experiments were run for time periods of 7, 14,

28, and 56 days. The vessels in the second set were run for periods of 14, 28,

56, 91, and 182 days. There were 66 vessels in each set of experiments.

8

Table 3. Composition of Topopah Spring tuff (whole rock)

from Hole USW-Gl, expressed as oxides.a

Species Weight %

SiO2 75.2

A12 03 12.4

K20 4.8

Na2O 3.1

Fe203 1.8

CaO 0.5

MgO 0.2

TiO2 0.1

Mn 0.06

Loss on ignition 1.0

99.16

a From R. A. Zielinski, Evaluation of Ash-flow Tuffs as Hosts forRadioactive Waste: Criteria Based on Selective Leaching of Manganese Oxides,U.S. Geological Survey, Denver, CO, Open File Report 83-480 (1983).

9

Table 4. Composition of "Equilibrated J-13" water at 200C

(first set of experiments).a

Species Weight ppm mm

Na

Si

Ca

K (estimated)

Mg

Al

B

Sr

Li

Fe

U

HCO3 (estimated)

-2so 4

NO3

C1

F

46.5

34.4

9.08

6.

0.96

0.63

0.16

0.045

0.044

0.01

0.0024

122.

17.3

7.60

7.15

2.4

2.02

1.22

0.23

0.15

0.040

0.023

0.015

0.0005

0.006

0.00018

0.00001

2.00

0.18

0.12

0.20

0.13

NO2

C-2'co 3

N2

02 1

not detected

1.9

0.57calculated

0.032

0.013

0.53

0.28

15.

9.1

a From J. K. Bates, D. F. Fischer, and T. J. Gerding, The Reaction of Glassin a Gamma Irradiated Saturated Tuff Environment. Part 1: SRL 165 Glass,Argonne National Laboratory, Argonne, IL, ANL-85-62 (1985).

10

Table 5. Parameters influencing composition of gas and liquid phases.

Parameter Value at 200C Value at 900C Reference

Vapor pressure of water (kPa) 2.339 70.14 62

Specific volume of saturated

liquid (m3/kg) 1.0018 x 10 3 1.036 x 103 62

Specific volume of saturated

vapor (m3/kg) 57.791 2.361 62

Henry's Law constant of

C°2 (M/Pa) 3.89 x 10 7 1 x 107 63,64

Apparent first dissociation

constant of 2CO3

(activity units) 4.1 x 10 7 4.0 x 10 7 44

(interpolated)

Second dissociation constant

of H2CO3 (activity units) 4.2 x 10-11 7.3 x 10 11 44

Ion product of water (M2) 6.8 x 10-15 3.5 x 10-13 65

Henry's Law constant of

N2 (M/Pa) 6.83 x 10 9 4.18 x 10 9 63

Henry's Law constant of

02 (M/Pa) 1.36 x 10-8 7.6 x 10-9 63

11

Table 6. Changes in compositions of gas and liquid phases due to heating,

in vessels originally containing 16 ml of water.

Temperature

200C 900C

Gas phase

Volume (m3 x 106) 5.4 4.85

Pressures (kPa)

N2

°2

Ar

Co 2

H2 0

Total

78.2

21.0

0.9

0.03

1.1 (50% R.H.)

101.23

106.5

28.6

1.3

0.42

70.1

206.92

Liquid phase

Volume (m3 x 106)

pH (at temperature)

Concentrations of species

that are affected (M)

CO2

HCO3

O-2Co3

N2

02

16.0

8.6 (calc.) 8-1 (meas.)

1.3 x 10 5

2.0 x 10 3

3.2 x 10 5

5.3 x 10-4

2.8 x 10-4

16.55

8.06 (calc.)

4.2 x 10 5

1.86 x 10 3

1.5 x 10 5

4.4 x 10 4

2.2 x 104

12

Table 7. Composition of standard dry air.a

Species Volume % Volume ppm

N2 78.084 + 0.004

02 20.946 + 0.002

co2 0.033 + 0.001

Ar 0.934 + 0.001

Ne 18.18 + 0.04

He 5.24 + 0.004

Kr 1.14 + 0.01

Xe 0.087 + 0.001

H2 0.5

CH4 2.

N2 0 0.5 0.1

a From R. C. Weast and M. J. Atle, Eds., CRC Handbook of Chemistry and

Physics, 63rd ed. (CRC Press, Boca Raton, FL, 1982).

13

The volumes of 'equilibrated J-13w water initially added to most of the

vessels ranged from 12.5 to 16.4 ml. In order to study the effect of changing

the air-water ratio, the water volumes in two vessels in each set of experi-

ments (G-113 and G-114 in the first set, and G-265 and G-266 in the second

set) ranged only from 4.6 to 6.2 ml.

In the first set of experiments, some vessels were found to be pressurized

to above atmospheric pressure after the irradiations, while in the second set

they were not. Most of the 182-day vessels in the second set were found to have

closure nuts that could be loosened by hand. To reduce damage to the gaskets,

the closures in the second set had not been tightened as much as those in the

first set, so later gasket creep may have loosened the seal somewhat. In two

experiments in the first set (numbers G-113 and G-114) and in three in the

second set (G-208, G-247, and G-248), the water was found to contain a

flocculent, rust-colored precipitate, and the stainless steel was visibly

corroded. In these cases, high levels of chloride ion and lower-than-normal pH

were also observed. Experiments G-113 and G-114 were run with deionized water,

while the others used equilibrated J-13" water. The solutions in G-208, G-247,

and G-248 were found to be depleted in dissolved NO- + NO- and S- compared to

initial values. In the second set of experiments, the solution was observed to

be yellow in several other vessels. Films resembling oil films were observed in

some vessels.

Analysis of the water was performed by inductively-coupled plasma optical

emission spectroscopy, ion chromatography, and atomic fluorescence. In the

second set, dissolved gas analysis was also performed, both by mass spectrometry

of frozen samples and by the Van Slyke method, which involves gas chromatography

of a stirred solution. The pH was measured for all solutions. Bates et al.

have published some of the data.1 8'1 9

14

3. DISCUSSION

3.1 OVERVIEW

In the experiments of Bates et al., two fluid media were in contact with

each other: the gas phase, which started out as air, and the liquid phase,

which started out in most cases as "equilibrated J-13a water having the

dissolved species shown in Table 4. As time progressed, interactions occurred

between the fluid phases and the other materials, namely the stainless steel,

the glass, and the tuff. From a radiation chemistry standpoint, this is a

very complex system. Radiation chemical experiments are usually performed on

highly purified systems with a minimum number of constituents in essentially

inert vessels. A precise theoretical analysis of this system would require a

time-dependent computer model incorporating at least two compartments to

represent the two fluid phases. Within each compartment, provision would need

to be made for inputting the yields of the primary radiolytic species and

calculating their reactions by means of coupled rate equations. The

significant reactions and their rates would have to be known for both phases

at the temperature of interest. Provisions would have to be made for

transport of species between the two phases, and the equations governing such

transport would have to be supplied. Significant interactions between the

fluid and solid phases would also have to be understood well enough to be

modeled mathematically.

At present, the capability to do all of this is not available, so far as

I know. Researchers at Chalk River have made progress on understanding the

irradiation of dry air and have published results of computer simulations.2 4 -26

Tokunaga and Suzuki have qualitatively described the radiation chemistry of

moist N2-02 mixtures.27 Busi et al. have presented a computer model for

mixtures of N2, 2' H20, and CO2.28 To my knowledge, however, modeling of the

radiation chemistry of a gaseous system incorporating significant proportions

of N2, 02, H20, and H2 (from radiolysis), such as was present in the Bates et

al. experiments, has not been performed.

Likewise, computer models of liquid water systems have been constructed

by researchers from Argonne,29 Harwell,30 Chalk River,31 Riso,32 Battelle

Pacific Northwest Laboratories,33 and Battelle Columbus Laboratories.34

However, these models have incorporated only a few solutes. To my knowledge,

15

modeling has not been performed on a bicarbonate water, although Swedish

researchers have discussed the importance of bicarbonate as a scavenger in

irradiated groundwater.35

Provision for transport of H2 and 2 between a gas and a liquid phase has

been made by some of these researchers. I am not aware of any computer models

that have incorporated transport of oxides of nitrogen over time. Burns et

al. have modeled the corrosion behavior of stainless and mild steels in the

presence of irradiated air and demineralized or distilled water.3 6 However,

to my knowledge, interactions with nuclear waste glass and rock have not been

modeled in a radiation chemistry code.

Since a thoroughgoing computer model of the system of interest was not

available, I elected to perform an approximate analysis of the radiation

chemistry, considering only the dominant reactions. In this report, I first

present an overview of the major features and the expected observable

results. I next consider the radiation chemistry in greater detail, beginning

with the fluid phases in isolation. I then consider transport of species

between the fluid phases and important interactions with the solid phases.

With this background, I calculate specific observable parameter changes and

compare them with the results of Bates et al. for cases in which corresponding

measurements were made.

3.2 CHANGES IN COMPOSITION DUE TO HEATING

In this paper I take the irradiation temperature to be 900C. Differences

in behavior at 920C, as actually used in the first set of experiments by Bates

et al., are expected to be minor. When a closed system containing air and an

aqueous solution such as equilibrated J-13" water is heated from room

temperature to 900C, a number of changes occur that affect the composition of

the two phases. The important parameters and their values at 200C and 900C

were given in Table 5. The main effects of heating on the vessels that

originally contained 16 ml of water and 5.4 ml of air are shown in Table 6.

Note that: water evaporates to produce a higher vapor pressure; the remaining

liquid water expands and reduces the gas space; a small amount of CO2 moves

from the liquid to the gas phase because of its reduced solubility; the

bicarbonate and carbonate concentrations drop slightly to maintain

equilibrium; the dissolved CO2 concentration rises slightly because of the

16

increased vapor pressures and the N2 and 02 concentrations drop slightly

because of decreased solubility. My calculations of these values assumed

constant concentrations for cations other than H+. I considered this valid

because the water was previously equilibrated" at 900C. Ionization of

silicic acid was not included; its effect would be to lower the pH only

slightly. It should be noted that a small amount of gas leakage may have

occurred from the 182-day vessels in the second set of experiments by Bates et

al. This would have had the effect of reducing the relative concentrations of

N2 2 A 02, and later 2 , with respect to water vapor, in the gas phase.

In the following discussion I focus first on the so-called blank" exper-

iments that started with 16 ml of water and did not contain rock or glass.

3.3 OVERVIEW OF MAJOR FEATURES AND EXPECTED OBSERVABLE RESULTS

Past research has shown that the stable products of low-dose-rate

irradiation of moist air are gaseous nitric acid (EN03), nitrous oxide N20),

and a small amount of ozone (03).6 In a liquid water phase, the primary

products of irradiation are OH, e, H3Ot, H202, H, H2, and HO2 (Table 8).

If three conditions are met (i.e., if the water is of high purity, the

container is inert and sealed, and only gamma irradiation is used), then the

free radical products H, OH, eaq, and 2 will react with the molecular

products H2 and H202 in such a way that the result is small steadystate

concentrations of H202 , H2, and 02 dissolved in the water, and there is no net

decomposition of the water as a function of time. However, in the Bates et

al. experiments there were several initially-dissolved species that have

significant reaction rates with the primary free radical products. In

addition, the species produced by irradiation of the gas phase would dissolve

in the liquid phase, and could then also interact with the primary free

radicals. Interactions would also take place between the free radicals and

the stainless steel vessels. It appears from previous investigations that

there would be four significant results of these processes:

1. A net decomposition (radiolysis) of water into H2 and 02 would occur as a

function of time.

2. Nitrogen would be fixed and added to the liquid phase, appearing at the

end of the experiments as nitrate (N0) and nitrite (N0) ions.

17

Table 8. Primary products of gamma irradiation of

liquid water for pH 5 to 9.a

Product G

H20 -4.08

OH (hydroxyl radical) 2.72

eaq (hydrated electron) 2.63

R30+ (hydronium ion) 2.63

H202 (hydrogen peroxide) 0.68

H (hydrogen atom) 0.55

H2 (hydrogen molecule) 0.45

HO2 (perhydroxyl radical) 0.026

a From J. W. T. Spinks and R. J. Woods, An Introduction to Radiation

Chemistry (Wiley, New York, 1976), 2nd ed.

18

3. An equivalent amount of hydrogen ions would be produced in the solution,

tending to lower the pH against the buffering action of the dissolved

bicarbonate as well as that of the rock and glass, where present.

4. Colloidal Fe(III) iron-containing substances would become suspended in

the solution as a result of vessel wall reactions.

3.4 GAS PHASE

If we use the gas composition at 900C in Table 7 and assume that the

gamma rays would interact with the species listed in proportion to their

fractions of the total electron density (a valid assumption for Compton

interactions, which are dominant here), we find that the energy should be

absorbed in about the following proportions:

N2 57%, H2 0 - 26%, and 02 17%, other gases negligible.

The production rates of the initial radiolytic products (ions, free

radicals, excited molecules, and electrons) can be determined by weighting the

yields, G for the pure gases by these proportions. The yield (in radiation

chemistry) is defined as the number of molecules of a particular species that

are created or destroyed when 100 eV of energy is absorbed from radiation. GX

represents the primary yield of species X, i.e., the yield evaluated before

any subsequent reactions take place. G(X) represents the final or observed

yield of species X in an experiment in which reactions may have occurred after

the initial production of the species. G(X) is measured, and G is inferred

or measured in experiments designed such that GX will be equal to G(X).

Willis and Boyd have summarized available data to evaluate the yields for

the various species formed in irradiation of inorganic gases.25 For the gases

that are important here, the species include N +, N+, N N + +,inclde 2, N N2, 2 0,O* + + + *22 2

°2 , H20, H, OH , H, OH, 20 H2 , and e . Some of these species are

produced in more than one energy state, so the total description of the system

becomes quite complex.

The ionic species produced in the irradiation would undergo a series of

reactions with other species present. 2 7 ' 2 8 The results would be additional

free radical and molecular species. The free radicals in turn would interact

with each other and with other species to produce HN03, N20, and 03, as

observed by Jones.6 Intermediate species in this process are thought to'

include NO, N02, and possibly N204. (NO3 and N205 are believed to be formed

19

in irradiation of dry air, but not moist air.) Further behavior is affected

by interaction with the liquid phase.

3.5 LIQUID PHASE

For aqueous solutions with total solute concentrations less than about

0.1 M, interactions of the radiation only with the water need to be

considered.3 7 When gamma rays interact with water, they excite electrons and

produce ionization by the photoelectric and Compton effects. The recoiling

electrons deposit their energy by interacting with electrons in other

molecules. The result is that excitation and ionization are produced in

concentrated regions called spurs and tracks. This results in formation

within these regions of the species H20+, e , and H20*. Reaction of H20+ with

water molecules then produces H30 ions and OH. Dissociation of excited

molecules produces H, OH, H2, and 0. Some electrons are captured by H30+

ions, producing H20 and H. Some of the caged H and OH radicals recombine to

form water. The electrons become solvated, forming eaq. All of the above

events occur within about 10-li seconds of the initial interaction of the

gamma ray with the water.

At about 10 10 seconds, the free radical species can diffuse far enough

to undergo significant numbers of reactions with each other. This occurs

first within the spurs and tracks, because of proximity. These reactions

produce H20, H2, and H202 within the spurs and tracks within a total elapsed

time of 10-8 seconds. So long as the concentrations of reactive solutes in

the water are less than about 103 M, the solutes will not significantly

interfere with the above processes.3 7 As can be seen from Tables 4 and 6, the

water in these experiments satisfies this criterion. At this stage, the

products of the irradiation are called primary products,' and their yields

are as shown in Table 8 for gamma irradiation of liquid water at room

temperature in the pH range from about 5 to 9. At 900C, the values might

differ by as much as 10%, as deduced from the work of Hochanadel and

Ghormley.3 8

During the period between 10-6 sec and about 1 sec after an initial gamma

interaction, the free radicals (OH, H, e , and HO2) and the molecular species

(H2 and H202) are able to diffuse from spurs and tracks into the bulk of the

solution. There the free radicals interact with each other, with the

20

molecular species, or with reactive solutes, if present. In the case of gamma

irradiation of very high purity (triply distilled) water in an inert, closed

container, small concentrations of 20 2, H2, and 02 build up, the latter from

decomposition of H202. Since there are insignificant concentrations of

solutes, the free radicals are able to react with the molecular species to

reform water, and there is no net decomposition of water over time.

When reactive solutes are present in the water, as is the case with

'equilibrated J-13R water, there is a competition for reaction with the free

radicals. They undergo second order reactions, the rates of which are given

by equations of the form

Rate moles k liter [R] moles x molesliter-sec' moles-sec liter 2] liter (Eq.l)

where [R1] is the concentration of free radical R, and [R2] is the

concentration of another reactive species, either another free radical, a

molecular product, or a solute.

In order to predict what the dominant reactions of each free radical will

be, we need to know the concentrations of these other species and the reaction

rate constants. The rate'constants for the significant reactions that take

place in pure water and common aqueous solutions have been measured and

tabulated. The list given in Table 9 summarizes the reactions and rate

constants presented by several authors. As can be seen, there is essential

agreement between authors. The values shown are for 250C. The free radical

reaction rates are not expected to be significantly different at 900C, because

the free radicals in water radiolysis are highly reactive, having low energies

of activation for their reactions.7'29 The reactions tend to be diffusion

limited, rather than activation limited. The p values for H20, 2, and H202

do shift with temperature. The 20 pK shift is the only significant one for

our purposes, and it has been taken into account.

A measure of the relative importance of solutes compared to free radicals

in reactions with free radicals can be obtained from a knowledge of the

maximum concentrations that could be reached by the free radicals during a

continuous irradiation of pure water, i.e., in the absence of reactive

solutes. These concentrations are fixed by the reactions of free radicals

with free radicals. For a maximum dose rate of 2 x 105 Mrad/hour, as used in

these experiments, the upper limit to the free radical concentrations is in

the range of 109 M.37 Of the solutes found in "equilibrated J-130 water in

21

Table 9. Reactions and rate constants for pure water at 250C.

Rate Constant (M-1 sec-1 or sec l)a

Reaction Spinks 40d Boyd Gordon BurnsWoods et al.3 1 et al.2 9 et al.3 6

1. eaq + H 2 O H + OH

2. ea + H+- -Haq3. eaq + OH - OH

4. ea + H 2 OH + OH

5. eaq + HO2 HO

6. aq +2 °2

7. (2 H+) ea + eaq aq- H2 + 2 OH

8. (H2O+) e + H

- H2 + OH

9. (H2 0+) eq + H°

- OH + 2 OH

10. (H2O+) eq + 2

- OH + HO2

11. OH + OH - H202

12. OH + HO2 02 + H20

13. OH + O2 02 + OH

14. OH + H - H20

15. OH + H2 - H + H20

16. OH + H202 HO2 + H20

17. OH + OH - H20 + 0

18. OH + HO- HO + OH19. H + H - H2

20. H + H202 - OH + H20

21. H + HO2 H202

22. H + O HO-

23. H + 02 - HO2

24. H + OH ea + H20aq 2

25. HO2 - H + 02

26. HO + 02 + HO2

16/-

2.35 x 1010

3.0 x 1010

1.2 x 1010

1.9 x 1010

5.4 x 109

2.5 x 1010

20/- 16/-

2.2

3.0

1.2

x 1010

x 1010

x 1010

2.3

3.0

1.3

x 1010

x 1010

x 1010

1.9 x 1010

5.0 x 109

2.5 x 1010

1.9 x 1010

5. x 109

2.5 x 1010

16/2.0 x 107

2.4 x 1010

3.0 x 1010

1.3 x 1010

2.0 x 1010

1.9 x 1 0 10

5.0 x 109

2.5 x 1010

3.5 x 1093.5 x 109 b

1.3 x 01 0

5.3 x 109

9. x 109

1.01 x 101

2.0 x 1010

4.9 x 107

2.7 x 107

1.2 x 1010

1.3

9.

2.0

x 1010

x 107

x 1010

1.3

5.5

1.2

9.0

2.0

3.6

3.3

1.2

5.0

1.0

9.0

2.0

2.0

1.8

2.1

8.0

4.5

8.9

x 1010

x 109

x 101

x 109

x i01 0

x 107

x 107

x 1010

x 1ix i010

x 107

x 1010

x iolo

x 1010

x 107

x 105/

x i010

x 107

1.

9.

2.

5.0 x 109

1. x 1010

1.5 x 1010

4.0 x 107

2.25 x 107

x 1010

x 107

x 1010

1.8

4.5

1.2

1.2

2.0

4.5

4.5

1.0

9.0

2.0

2.0

1.9

8.5

5.0

1.5

x 108

x 109

x 1010

x 1010

x 1010

x 107

x 107

x 1010

x 107

x 1010

x 1010

x 1010

x 105/

x 10 10

x 107

1.9 x 1010

2.3 x 107

pK = 4.88

4.4 x 107

1.9 x 1010

1.5 x 107

p = 4.88

22

Table 9. Continued.

27. HO2 + O2 .. H202 + 2

28. (2 H20+) 2 + 2

O 02 + 202 + 2 OH

29. H+ + 0H = 20

2.5 x 106 2.0 x 106 1.6(for

x 106pH=6.5)

2.7 x 106

0

1.43 x 1011/ 1.43 x 1011/

2.6 x 10-5

0

1.438

2.6 x

1.0 x

1.022

2.0 x

30. H + H 202 O + 2 °

x 1011/

10-5

io8,

x 104

1010/31. H + HO H 2 02

32.

33.

34.

35.

36.

37.

38.

39.

o + 2 - O + OH (0.17 to 2) x

o + 2 - 03 2.9 x 109

O + H 2 - H + OH 8. x 107

o + 202 H2 + 2 5.3 x 108

H2 + O 0H + 2

3 + H2 °2 °2 + 2 + H 20 -

0 + HO- + 2 +OH -

3- 02 +

2.0 x 1010/

3.56 x 10-2

107 1.7 x 106

3.0 x 109

8.0 x 107

2.0 x 108

8.0 x 108

1.6 x 106

8.9 x 105

3.0 x 102(pH - 13)

2.5 x 105

6.0 x 102(pH = 13)

40. O +

41. 2 -

H2 ` 2 + H + OH

products

42. 2 + 2 2-

43. OH= 0 + H+

+ 02 5.0x 10

pK . 11.9

NOTE: Where two values are shown, the first is for the forward reaction and thesecond is for the reverse reaction. -a Second order rates are given as M 1 sec 1 first order rates as sec-1, for k,

not 2 k.b Reaction is actually given as e + HO - 0 + oH-.

23

significant concentrations (Table 4), HCO3, NO-, Cl, CO 2 , co and 2 are

known to have significant reaction rates (i.e, greater than 107 M 1 sec'1)

with one or another of the free radical species. Since all these solutes have

concentrations greater than 105 M, they can clearly dominate the

free-radical/free-radical reactions in competing for free radicals.

In addition to reacting with dissolved solutes, the free radicals can

also react with suspended colloids. Burns et al. found that Fe(III) iron

oxide colloid formed upon gamma irradiation of distilled or deionized water

and air in stainless steel vessels, and they believe that reactions between

the free radicals and this colloid gave rise to the decomposition of water

they observed.36

In the following discussion, I consider each of the free radical species

in turn and determine what its dominant initial reactions with the solutes

would be in the solution (described in Tables 4 and 6) using measured reaction

rates reported in the literature. (Note that several solutes are present from

the beginning, whereas the colloids form during irradiation.)

As can be seen in Table 8, the hydroxyl radicals (OH) are the products

formed with the highest yield. Their initial reactions would primarily be

with chloride ions:

OH + C - ClOH k4 4 =4.3 x 109 M 1 sec&1 (see Ref. 39),

C1OH + H+ C + H0 k45 = 2.1 x 1010 M 1 sec 1 (see Ref. 39),

C1 + C - C k46 = 2.1 x 1010 M 1 sec 1 (see Ref. 39).24[f be2460 e~The product k44[Cl would be 2 x 105 sec 1 initially.

The hydroxyl radicals (OH) would also react with bicarbonate and

carbonate ions:

OH + HCO CO- + H k47 = 1.25 x 107 M 1 sec 1 (see Ref. 37),OH+CO 23C3+O2 k41xl 8 IC

OH + C2 CO + OH- k48 4.1 x 108 M1 sec1 (see Ref. 37).

The product k47 [HCO] is 2.3 x 104 sec . For C03 , the product is36.1 x 10 . The carbonate radical C03 reacts with the superoxide ion 02 (the

origin of which is discussed below), and the product is thought to be CO-2

(see Ref. 40):

co + - co-2 k49 = (1.5 to 15) x 108 M 1 sec& 1 (see Refs. 40-43).

This species in turn is thought to decay by a first order process:

co;2 - CO;2 + 02 (see Ref. 40) (Reaction 50).

The carbonate ion would then associate to form bicarbonate at pH 8:

HCO_ : H+ + CO2 pK51 5 10.1 at 90 (see Ref. 44).

24

Since the 02 has originated as 02 (see below), and one 30 ion is formed as a

primary product for each e radical formed in the irradiation, there would beaqno net chemical change from this sequence.

At the beginning of the experiments, there was no measurable nitrite in

the solution, but some was found at the end, so I will consider its

reactions. When nitrite was present, it would also have had a significant

reaction rate with OH radicals:

OH + N- N 2 +OH20 x 109 M 1 sec-1 (see Ref. 37).

At the end of the blanku experiments that received the highest doses, the

nitrite concentration was 2.1 x 104 M. The product k52[NO-] was then

1.3 x 106 sec1l, which means this reaction would dominate those mentioned

above after N had built up. The NO2 thus formed would react with water to

produce nitrate and nitrite45:

2 N2 +~ ° 0 2 + + N-_ + N (Reaction 53).

As the irradiation proceeded, H2, formed as a primary product, would

begin to build up and would compete with these other species for reaction with

OH radicals. It would continue to build up until its rate of reaction with

OH equaled its rate of formation. To achieve steadystate concentration, the

E2 would have to react with about one-sixth of the OH radicals, because G is

about one-sixth of GH (see Table 8). To have had a reaction rate equal

to one-fifth of the sum of the reaction rates of C- and HCO3, which were the

competitors at the beginning of the experiment, the H2 concentration would

have had to have been 9.1 x 10-4 M, which requires an H2 pressure above the

solution of 130 kPa. Such a pressure would not have been reached before the

N0 concentration was observed to rise significantly, so H2 would have

continued to build up to higher pressures as competition increased for the

OH. To have had a reaction rate equal to one-fifth that of the

NO2, C, and HCO3, which were the solute competitors at the end of the

experiment, the 2 concentration would have had to have been 6 x 10 3 M, which

would have required an 2 pressure above the solution of about 850 kPa at

900C. This means that relatively high H2 pressures would have been required

before the H2 pressure would have leveled off. For comparison, Gray and

Simonson found that gamma radiolysis of Permian Basin saturated brine at 750C

produced a total steadystate gas pressure (H2 plus 02) of about 10 MPa.46

This is a much more concentrated solution than equilibrated J-13N water.

25

In their experiments mentioned earlier, Burns et al. found that gamma

irradiation of air-saturated, triply distilled or demineralized water in

stainless steel vessels at 30 to 50C produced total steadystate pressures of

300 to 450 kPa, including H2, 2, and N2.3 6 As noted, they invoked the

presence of a colloid to explain the results. If colloid-free radical

reactions were significant in comparison to solute-free radical reactions at

the end of the the Bates et al. experiments, then the H2 pressure required for

steadystate would have been significantly higher than 850 kPa.

The radical species produced with the next highest yield is the hydrated

electron (eq). In oxygenated solution, superoxide ions are formed:

e + ° 0 ° = 1.9 x 101 M 1 sec 1 (see Ref. 37).aq 2 I6 1.x10 M -In the Bates et al. experiments that had 16 ml of water at the outset, the

initial concentration of 02 in the liquid phase would have been about

2.2 x 10 4 M (Table 6). The product k6[021 would then have been about

4 x 106 sec 1 for this reaction.

Competing reactions would have been those with nitrate, dissolved C02,

chlorine atoms, and ferric ions:

eaq + NO . NO 2 (- NO2 + 2 OH) k54 10 M1 sec 1 (see Ref. 37),

eaq + C2 02 k55 = 7.7 x 109 1 sec 1 (see Ref. 37),

eq + C1 - C k 6 + 1.0 x 109 M 1 sec& 1 (see Ref. 7),

eaq + Fe+3- Fe+2 k57 = 5 x 1010 M 1 sec 1 (see Ref. 37).

Using the initial N concentration (Table 4), k5 4[N0] would have been

1.3 x 106 sec 1, and k5 5[CO2] would have been about 3.2 x 105 sec 1 . The

very low Fe+3 concentration would have made it a negligible contributor at the

beginning of the experiments, and the pH buffering would have kept its

concentration low. It is difficult to calculate the steadystate C1

concentration, but the maximum possible value of k5 6[Cl] if only the initially

dissolved C is considered would have been 2 x 105 sec- . This indicates

that nearly 70% of the eaq would have reacted with 02, about 22% would have

reacted with NO3, and about 6% would have reacted with C02 at the outset. As

the irradiation proceeded, the 02 would have built up, taking a larger share

of the eaq reactions, and colloids also could have become important reactants.

As discussed above, the 2 produced would react with carbonate radicals,

producing 02. It would also react with OH by Reaction 13, again producing 02:

OH = 2-- 0H + 2 k 3 = 1.0 x 1010 1 sec 1 (see Ref. 37).

Burns et al. predict that the 02 would also react significantly with the

stainless steel vessel walls.3 6

26

The HO2 radicals produced as primary species would dissociate to °2 (at

pH above about 5) and would react in these same ways. The NO2 produced would

react with water to form NO- and NO-, as discussed earlier (Reaction 53).

The carboxyl radicals (CO2) would remain dissociated at pH 8:

COOH- H+ + C02 pK58 = 3.9 + 0.3 (see Ref. 47).

In oxygenated solution, the charge transfer reaction

CO° + 02 - C 2 + 2 (Reaction 59)

has been found to be fast,47 and for [02] = 10 4 M, it would overwhelm the

dimerization reaction to form oxalate:

-72 2 x sec (see Ref. 48)COw2 2O ' 20 k60 lO 1 109 -so that very little oxalate would be formed. Furthermore, in solutions

containing 021 oxalate is destroyed by reactions with OH radicals

C202 + OH - C20 + H k61 = 6 x 106 M1 sec 1 (see Ref. 49)

which are followed by reactions with oxygen:

C20i + 02 - °2 + 2 CO2 (Reaction 62) (see Ref. 49).

In the presence of Ca++ ions, as in "equilibrated J-13" water, oxalate

(if present) would precipitate as calcium oxalate, which has a solubility of

about 104 M (see Ref. 50). The fact that the observed Ca++ concentration in

solution did not decrease from its original value of 2.3 x 10 4 M supports the

contention that little oxalate was formed. In the experiments of Hasselstrom

and Henry, oxalic acid was produced by irradiation of bicarbonate and,

necessarily, CO2 solutions.51 Apparently there was insufficient oxygen to

oxidize the carboxyl radicals, so that the dimerization to form oxalate was

free to proceed. Their solutions were irradiated in polyethylene bags, and

they did not perform gas analysis.

If the CO- became associated to form COOH, it could react to form formate

ions or formic acid47:

COOS + CO2 - CO2 + HCOO (Reaction 63)

or COOH + COOH - HCOOH + C 2 (Reaction 64).

The formate and formic acid thus formed would remain soluble. Production of

formate is not likely in the bulk of the liquid phase, because the pH there

was buffered near 8, and the C would not associate. However, it might have

occurred in the liquid film around the top of the gas space, where nitric acid

production and deposition could drive the pH down in the absence of

buffering. It should be noted that past experiments in which formate was

produced from dissolved CO2 were either oxygen-free8 10 or were driven to low

27

pH by simultaneous formation of nitric acid.11'12 Either of these conditions

would prevent the carboxyl radicals from transferring charge to oxygen, and

thus make them available to form formate or oxalate.

The other primary free radical produced in the irradiation is the

hydrogen atom. At the outset, its only significant reactions would be with

dissolved oxygen and with Cl produced in the OH reaction mentioned earlier:

H + °2 HO2 = 1.9 x 1010 M 1 sec 1 (see Ref. 37),

H + C1 - H + C1 k65 = 5.0 x 109 M 1 sec 1 (see Ref. 7).

The HO2 would dissociate to °2 and react as described earlier to produce 2*

The products k23[02] and k65[Cl] would be 4.2 x 106 sec 1 and 1 x 106 sec 1

(maximum), respectively. As the irradiation proceeded, nitrite would be

formed as discussed earlier, and H atoms could react with it:

H+ N2 - NO + OH k66 = x 109 M 1 sec 1 (see Ref. 37).

At the end of the highest dose experiments, k66 [N02] was about 105 sec 1 for

this reaction. In the Bates et al. experiments, the 2 concentration as a

function of time probably represents a balance between production by water

radiolysis and loss to reaction with the stainless steel. Data from corrosion

experiments indicate that reaction with the metal in these experiments would

account for only a small part of the 02 produced by radiolysis.52 This result

was also found by Burns et al.36 It therefore appears that the 02 concentra-

tion would rise, and 02 would dominate the H reactions during the entire experi-

ment. The NO probably would react rapidly with an oxidizing species in the

water to again form NO2 and then would react with water to form NO- and NO3.

Having discussed the fates of the primary free radical species, we turn now

to the molecular species H2 and H202. As discussed already, H2 can be expected

to rise in pressure to at least 850 kPa if the irradiation proceeds for a time

long enough to reach steadystate conditions. In pure water, the H202

concentration is limited by reactions with ei H, and OH, as noted above.

Where these were effectively scavenged by solutes, H202 could be expected to

build up. We have observed this in room temperature irradiation corrosion

experiments that included only a glass vessel, J-13 water, and a stainless steel

electrode.53 However, since H202 is thermodynamically unstable at all

temperatures, it would thermally decompose at elevated temperatures and in the

presence of materials that are catalytic for the decomposition reaction

2H202 - 2 H20 + 02 (Reaction 67).

28

Dissolved Fe+3 is known to be an effective catalyst, producing measurable

effects at concentrations less than 0.1 ppm.5 Solid materials found in the

tuff and glass may also catalyze this decomposition. Manganese oxide, which is

found in the tuff, is one example of a known effective catalyst.54 H202 could

also react with Cl:

H22 + C1 -H+ + C + 2 k 6 8 5.0 x 109 M 1 sec 1 (see Ref. 7).

It seems probable that the 202 concentrations in these experiments were limited

to small values by these reactions. The decomposition of 202 would give rise

to oxygen gas, and the amount would be essentially one-half the amount of

hydrogen given off, since the stoichiometry of the solution must continue to

correspond closely to 20.

3.6 GAS AND LIQUID PHASES CONSIDERED TOGETHER

Because gas and liquid phases are in contact in the experiments of Bates et

al., we must consider exchange of gaseous species between them during the irradi-

ation. As noted in Section 3.4, the stable products formed during irradiation of

the gas phase would be gaseous N03, N20, and 03. In addition, there would be

transient species such as NO, NO2, and probably N204 that could conceivably

dissolve in the liquid phase before they reacted to form HNO3 or N20. To

determine which of these processes would be more rapid, we would need to calcu-

late the lifetimes of these transient species by modeling the gas phase reaction

sequence and then compare them to typical diffusion times to reach the liquid

phase in the geometry used. This has not been done in this analysis because of

the complexity of the system and the uncertainty of some aspects of the reaction

sequence. However, even though the detailed mechanism is not established, exper-

imental measurements have been performed on the two-phase air-water system.5'55

Burns et al.7 have derived an equation describing the fixation of nitrogen

as a function of time and other parameters for a sealed system:

N =2 CR [1 - exp (-1.45 x 105 GDt)] (Eq. 2)

where N is the concentration of ENO3 in the water in moles/liter at time t,

Co is the initial concentration of N2 in the air in moles/liter,

R is the ratio of the volume of air to the volume of water,

G is the yield ( 1.9) (using arwell data as interpreted by Burns et al.7),

D is the dose rate in rad/hr, and

t is the time in hours.

29

For total doses much less than 3.6 x 104 Mrad, this equation reduces to

N = 2.90 x 10 5 CO RGDt. (Eq. 3)

In the Bates et al. experiments, the maximum dose was 269 Mrad, so this form

of the equation is adequate. As can be seen, the production of fixed nitrogen

is directly proportional to R and to the dose Dt.

Although Equation 2 appears to have been derived using unbuffered,

initially neutral water, I have applied it to the Bates et al. data on the

grounds that the total fixed nitrogen concentration does not appear to be very

sensitive to pH under oxidizing conditions.55 This probably results from the

facts that the oxidation occurs in the gas phase and that once the nitrogen is

fixed as nitrate or nitrite in the solution, it tends to remain in the

solution, because to escape, it would have to be reduced to NO, and this

species is rapidly oxidized in an irradiated solution regardless of the pH,

forming NO2 and then NO2 and NO3. The movement of fixed nitrogen is thus a

one-way process under these conditions.

The N20 produced in the gas phase may dissolve in the liquid phase and be

destroyed by reactions with eag:

eaq + N20 - N2 + 0 (+ H - OH) k69 = 8.7 x 109 M41 sec 1 (see Ref. 38).

Ozone rapidly decomposes upon dissolution in water at high temperatures and

neutral or high pH. The mechanism is not completely understood, but the

formation of OH radicals is a significant aspect of it.56

The gases generated in the liquid phase in excess of their solubilities

must move to the gas phase. These include H2, 02, and the N2 resulting from

the above N20 reaction.

3.7 INTERACTIONS WITH THE SOLID PHASE

As mentioned earlier, we expect that some oxidation of the stainless

steel will occur, although it should be limited by the formation of a

protective layer under the pH and redox conditions of most of these

experiments. Dissolution of iron and chromium should be slight, so long as

the pH remains near 8.57 Nickel and manganese do have some solubility under

these conditions, but their dissolution should be limited by the protective

layer.

Burns et al. have gamma irradiated sealed cylinders made of Types 304 and

321 stainless steel, containing air and either triply distilled or

30

demineralized water.36 Temperatures in their experiments ranged between 30

and 500C, and dose rates were between 0.15 and 2 rad/hour, with doses

extending up to 1000 Mrad. They analyzed the gases inside the cylinders and

monitored their pressures. The amount of corrosion occurring was calculated

from the deficit in oxygen gas observed compared to one-half the observed

hydrogen. They explained their gas composition results on the basis of the

formation of an Fe(III) iron-oxide colloid, which they observed at the end of

some of their experiments.

As noted in the description of the Bates et al. results in Section 2,

yellow solutions were found in several vessels, and rusty precipitates were

found in others. It seems clear that an iron-containing colloid was formed in

the Bates et al. experiments as well. Flocculation of the colloid may account

for the anion depletion.

Stainless steel is not a very efficient catalyst for the decomposition of

H202. The stainless steel surface could act as a substrate for precipitation

of species from the solution that may have dissolved from the glass or tuff

and later become less soluble because of changing solution conditions or

changes in the chemical form or oxidation state of the species. The stainless

steel could also affect the nitrite-nitrate ratio in the solution, because of

the ability of iron to reduce nitrate to nitrite. In unirradiated open system

corrosion experiments with J-13 water and stainless steel, we have observed

the gradual disappearance of fixed nitrogen from the water completely, over a

period of a few months.58 We presume that this disappearance results from the

reduction of nitrate to nitrite and then to N20, which is a relatively inert

gas and could diffuse out of an open system.

The tuff rock would serve as a powerful pH buffer for the solution by

exchanging Na+, K+, or Ca+2 ions for H+ ions. This has been observed

extensively in the past with feldspar minerals, as mentioned in the

Introduction (Section 1). The manganese oxide and iron in the tuff could

serve as catalysts for B202 decomposition, as mentioned in Section 3.5.

Transient primary radical species formed near the surface of the tuff could

react with it directly, particularly with species such as iron, which has more

than one possible oxidation state. Dissolution of some minerals from the tuff

could be expected, as well as precipitation of others.

Reports by Bates and Oversbyle and Bates et al.19 describe the behavior

of the glass. This behavior includes dissolution and precipitation of various

31

species. As is true for the tuff, we expect that the iron and manganese oxide

in the glass could serve as decomposition catalysts for H202, and the iron

could also enter into reactions with primary free radical species.

In the Bates et al. experiments, the SRL A glass was doped with actinides

(as noted in Section 2), and the resulting alpha emission necessarily

delivered ionizing radiation to the liquid in contact with the glass

surface. However, approximate calculations indicate that the dose rate would

have been less than 103 rad/hour at the interface for the doping level used,

compared to the 104 and 2 x 105 rad/hr gamma dose rates absorbed by the

solutions in the experiments. When one considers in addition that only a

layer of liquid of the order of 10 3 cm thick received the alpha dose, it is

clear that the alpha contribution to the bulk irradiation of the solution was

small compared to the gamma contribution in these experiments. However, near

the surface of the glass in the lower dose rate experiments, the alpha

irradiation could have caused a significant increase in the concentrations of

the molecular products H2 and H202, since their yields are considerably higher

in alpha radiolysis than in gamma radiolysis.37 Bates noted that in both sets

of experiments the SRL A glass reacted slightly faster than the SRL U

glass.18,19,22 This may have been a consequence of alpha radiolysis.

3.8 DECOMPOSITION OF WATER

As has already been alluded to in Section 3.5, one effect of gamma

irradiation on an aqueous solution containing solutes that can react with the

primary free radical species is decomposition of water into H2 and 2

In the Bates et al. experiments, it appears that the solutes present

initially were in high enough concentration to prevent the radicals from

reacting significantly with the molecular products in the bulk of the solution,

but they would not have been so high as to interfere with the formation of the

molecular products in the tracks and spurs. It thus appears that the initial

water decomposition yield G(-H2 0) should have been approximately equal to GH

i.e., 0.45. Applying this value to the "blank" experiments irradiated at 2

2 x 105 rad/hr for the maximum time of 56 days, giving a total dose of 269

Mrad, the exposure of 16 grams of water would have led to the decomposition to

H2 and 02 of 3.6 x 10 2 grams, or 2 millimoles, assuming no back reactions

occurred. Using the ideal gas law and a gas space (at room temperature) of

32

5.4 ml, this amount of gas would have developed a pressure increase of over

1.3 MPa after cooling to room temperature. As noted in Section 3.5, the back

reaction of 2 with OH radicals would actually limit the 2 pressure to about

850 kPa at 900C, if reactions'with colloids are not considered. Adding an 02

pressure equal to half this value would produce a total of about 1.28 MPa at

900C. This would be equivalent to 1.03 MPa at 200C. This indicates that at

the highest dose in the Bates et al. experiments steady-state gas pressures

should have been achieved. The total pressure, including the original air,

would have been 1.13 MPa at 200C. Bates et al. observed pressurization of the

higher dose vessels (as noted in Section 2), although the pressures were not

measured.19

In their second set of experiments, Bates et al. applied a maximum dose

of about 43.7 Mrad, or less than one-sixth of that in the first set. This

would help to account for the lack of observed pressurization in the second

set. It should be noted also that some loss of gases through or around the

silicone rubber gasket may have occurred in this set, as mentioned in

Section 2, and some reaction of oxygen with the stainless steel also occurred;

both of these processes would have reduced the pressure. The time available

for these processes to occur was over three times as long for the longest

duration irradiations in the second set of experiments as for those in the

first set. Although calibrated dissolved gas analyses were not complete at

the time this paper was written, Bates reports that considerable amounts of

hydrogen were found (in addition to nitrogen, oxygen, and carbon dioxide) in

the second set of experiments, and that the relative amounts found were

consistent with the hypothesis that some gas leakage occurred in the second

set.22

3.9 FIXATION OF NITROGEN

As noted in Section 3.6, Equation 2 (Burns et al.7) predicts that the

amount of radiation-fixed nitrogen should be linear with dose in this range.

The equation also predicts that the amount should be proportional to R, the

ratio of gas to liquid volume. Since R varied between vessels, it is

necessary to correct for this in making comparisons. Before such a correction

can be applied, the value of the initial total fixed nitrogen concentration

must be subtracted from the measured fixed nitrogen data.

33

An initial plot of the data for the 2 x 105 rad/hour experiments showed

that there was considerable scatter and that subtraction of the reported

initial value (7.6 ppm or 1.22 x 10 M NO ) did not result in a best fit

line passing through the origin. Nevertheless, it was assumed that such a

line must obtain, and that the points at 28 and 56 days were the most

reliable, being highest above the initial value. Consequently, the value

chosen for the initial total fixed nitrogen concentration was that which

resulted in a straight line passing through the origin and the average values

at 28 and 56 days, uncorrected for R value. This turned out to be 1.65 x 10 4

M (or 10.2 ppm) NO-. This initial value was subtracted from the data, and the

correction for R was then applied.

The resulting data for the 2 x 105 rad/hour experiments are plotted in

Fig. 1, normalized to R = 0.337, which was the value at room temperature for

the vessels containing water and air only. The solid line represents the

prediction of the Burns et al. equation, indicating 1.6 x 10 4 M of additional

fixed nitrogen concentration for a dose of 269 Mrad. The agreement is

satisfactory, considering the uncertainties in the data and in the yield.

Each data point is an average from two replicate experiments. In some cases

replicates differed by more than 15% in reported total fixed nitrogen

concentrations.

Bates et al. give the precision of anion analyses as 5%, and the

accuracy also as +5%. The uncertainty in dose was +10%. The uncertainty in G

value was not given by Burns et al.,7 but a value of +15% would not be

unreasonable, based on the dose uncertainty in the reference they cite.55

Because of these uncertainties, it is not possible to determine whether there

were significant differences in nitrogen fixation rates among the various

vessels. All that can be said is that the rates were essentially as predicted

by the Burns et al. equation.

For the second set of experiments, the dose was less than one-sixth as

great as it was for the first set. The predicted increase in fixed nitrogen

concentration would thus be 2.6 x 10-5 moles/liter for those vessels having

low gas-to-liquid volume ratios. The observed total fixed nitrogen

concentration does not show a significant increase over the course of the

experiments, and the scatter in values is of the order of 1 to 3 x 10 5

moles/liter. As can be seen from Fig. 1, the predicted increase is less than

the precision of the data and would therefore not be expected to be seen.

34

1.80

1.60

Water and air

Burns et al.equation

//

/

Cl-

CL10

c

W-

%._

0c.x

C

1w

C0

.,

w2!e

LU

1.40

1.20

1.00

0.80

0.60

0.40

0.20

0

-0.20

//

/

/

-SRL U and tuff

-SRL U

-Crushed SRL A"-Crushed SRL U

-SRL A and tuff

//

/

/

/ //

// /

/ ///

/

-d

20 30 40 50 60Time (days)

-0.40

Figure 1. Nitrogen fixation in 2 X 105 rad/hr experiment.

35

3.10 HYDROGEN ION PRODUCTION

The production of nitrite and nitrate in the aqueous solution was

accompanied by an equimolar production of hydrogen ions. For the experiments

with 16 ml of water and a dose of 269 Mrads, the total added H would

therefore have amounted to 1.6 x 10-4 moles/liter. Since the initial

bicarbonate concentration was more than 10 times this level, we would expect

it to have buffered the added H and thus to have prevented a large pH

change. This is in fact what is observed in the Bates et al. data. By

contrast, experiments performed by others using deionized or distilled water

have shown shifts of pH into the 3-4 range.2 The changes in pH due to

nitrogen fixation for the second set of experiments by Bates et al. should be

even less significant, since the maximum dose was one-sixth as large.

3.11 NITRITE-NITRATE RATIO

Most past experiments involving irradiation of two-phase air-water

systems that have been reported in the literature have made use of deionized

or distilled water. In such cases the solution is unbuffered, and generation

of nitric acid drives the pH so low that any nitrite present tends to

associate with H to form nitrous acid, HNO2 (PKa = 5.22 at 250C).45 This

acid is thermally unstable and decomposes according the the reaction:

3 HNO2 H+ + N + 2 NO + H0 (Reaction 70) (see Ref. 59).

It also reacts with H202:

H202 + HNO2 - H + NO + H20 (Reaction 71) (see Ref. 59).

NO is rapidly oxidized in irradiated solution to NO2. The NO2 reacts with the

water to form equimolar amounts of NO- and NO3. The NO- then associates and

repeats the cycle. The result is that any NO- initially formed is converted

to NO-, and this is the only nitrogen species observed at the end of the

experiment.

Linacre and Marsh reported some experiments in which nitrogen gas and

water containing KOH were irradiated.55 In these cases, only nitrite was

observed in solution after irradiation. The explanation they gave was that

the nitrogen was fixed first as NO2, which did not associate because of the

alkaline pH, and was therefore stable. Conditions in these experiments would

have been less oxidizing than in the Bates et al. experiments, because of the

absence of oxygen at the start of the experiments.

36

The pH in most of the Bates et al. experiments was also buffered above

the p of HNO2. However, oxygen was present in addition to the nitrogen. A

balance would therefore have been achieved between NO2 and NO3 that reflected

the oxidation conditions in the solution. Figures 2 and 3 show the nitrite-

nitrate ratios for the two sets of experiments. It can be seen that the

higher dose rate (Fig. 3) generally produced lower ratios of nitrite to

nitrate. At the lower dose rate (Fig. 2), the ratios were lowest for the

vessels containing water and air only. In the higher rate experiments, those

with either tuff or crushed glass present had higher ratios than the others at

the end of the irradiation.

These results can probably be explained in part by the balance between

production of 202 in the solution by irradiation and its catalytic

decomposition by materials such as manganese oxides in the tuff and the

glass. The higher dose rate would have produced a higher steadystate H202

concentration. In the water-only vessels, there would have been less

catalytic material, so that higher 202 concentrations would have existed.

The vessels with either tuff or crushed glass would have had high catalytic

surface areas, tending to decompose the H202. Although H202 does not react

directly with NO; to produce NO;1 it can affect the relative concentrations of

ea, and OH radicals by the reaction:

ea + H2 02 OH + OH k4 = 1.3 x 1010 kH1 sec 1 (Ref. 37).

These two free radicals in turn can affect the nitrite-nitrate balance

according to the reactions:

(H20 +) N03 + eq - N 2 2 k72 1-1 x 1010 e-1 sec-1 (Ref. 3,

NO~~+OH-~NO2 +OH k 7 3 6.xl 9 1 sec-l (Ref. 37),N°2 + O - N02 + H k73 -6.2 x 109 M1sa1 (e.3)

2 N02 + H20 - N + NO + 2 + (Reaction 53).

In summary, the greater catalytic activity of the tuff and/or crushed

glass would reduce the 202 concentration. This in turn would raise the ratio

of eaq to OH concentrations. The result of this would be to raise the ratio

of N2 to NO3 concentrations. In addition, H202 would tend to oxidize any

trace organics present, thus decreasing the concentrations of these other

possible reducing agents. The oil-like films observed in some experiments may

be evidence for the presence of organics.

At the end of the lower dose rate experiments, it appears that the

nitrite-nitrate ratios were influenced strongly by the types of solid

materials present. SRL A glass produced higher ratios than SRL U glass, and

37

7'-

6

5 _

SRL A

SRL A and tuft

SRL U

SRL U and tuff

0z

i

I 0z

._0ir

4 _

3 _

2 _

1

Water and air only

0 50 100 150 200Time (days)

Figure 2. Nitrite-nitrate ratios in 104 rad/hr experiments.

38

Crushed SRL U2

I 0z-

I N0z

..P

1

tuff

Water andair only

I I I I I0

0 10 20 30

Time (days)

40 50 60

Figure 3. Nitrite-Nitrate ratios in 2 X 105 rad/hr experiments.

39

the tuff appeared to be lowering the ratios that these glasses would otherwise

produce. I do not currently understand the mechanisms behind this behavior.

Lower dose rate experiments now in progress should help us to understand this

behavior better.

3.12 LIKELY EFFECTS ON GLASS REACTION

Gamma irradiation has been found to affect glass dissolution in deionized

water in two ways: 1) indirectly through promotion of an acid pH, which

increases the solubility of some elements and makes H ions available for

attack of the glass structure and for ion exchange and 2) production of free

radicals such as OH, which can attack the glass directly. It appears that

both these effects would be somewhat ameliorated in equilibrated J=13"

water. The presence of bicarbonate in the solution serves to buffer the pH,

so that the solution remains near neutral. In addition, the initial presence

of C1 and HCO3, and NO2 after some irradiation, serves to lower the OH

concentration in solution. We would therefore not expect as large an effect

on glass dissolution from irradiation with this solution as is observed with

deionized or distilled water.

Bates has noted, however, that the solubilities of actinides are highly

pH-dependent in the pH range of these experiments, so that even the small

changes caused by irradiation may account for the observed differences in the

actinide behavior.2 2 Nash et al. have discussed the effects of irradiation on

actinide behavior during reaction of glass with high purity water.6 0 Bates et

al. are planning to publish a more complete discussion of the glass behavior

they observed in their experiments, at a later date.

3.13 HIGH GAS-TO-LIQUID VOLUME RATIO EXPERIMENTS

As noted in Section 2, four vessels were irradiated with a smaller

initial liquid volume, which produced a larger gas-to-liquid volume ratio. In

the first set of experiments, test no. G-113 began with 4.60 ml of deionized

water, while test no. G-114 began with 5.16 ml of deionized water. These two

vessels also contained a stainless steel waste form holder, but no tuff or

glass. They were irradiated for 28 days at 90 C and a dose rate of 2 x 105

rad/hour, producing a total dose of 134 Mrad. At the end of the irradiation,

40

the vessels were found to contain a flocculent, rst-colored precipitate, and

the vessels and waste form holders were visibly corroded. The solutions were

very acidic, with the p values being 1.82 and 2.58 for G-113 and G-114,

respectively. The solutions in both vessels contained significant amounts of

elements known to be in stainless steel (Cr, Cu, Fe, Mn, i, and Si). They

also contained 27.2 and 25.6 ppm of nitrate, respectively, and no detectable

nitrite. The solutions also contained 589 and 103 ppm of chloride,

respectively. The precipitate was found to consist of amorphous material,

high in Fe, Cr, and Si.

Tests G-265 and G-266 in the second set of experiments had comparable

amounts of equilibrated J-130 (rather than deionized) water. They were

irradiated for 182 days at 900C and a dose rate of 1 x 104 rad/hr, producing a

total dose of 44 Mrad. They were found to have pH values of 6.94 and 6.88,

respectively, and normal equilibrated J-13" water concentrations of the-2

anions F, C, and S . The nitrate concentrations were 4.4 and 2.1 ppm,4respectively, and the nitrite concentrations were 10.8 and 7.5 ppm,

respectively. No corrosion or rust-colored precipitate was evident. A small

amount of formate was detected in the solutions.

In test G-113, the predicted concentration of nitrate from Equation 2

(Burns et al.7) is 8.3 x 10 4 M, taking account of the water expansion and

contraction and the space occupied by the holder. The observed concentration

was 4.4 x 10 4 M, almost a factor of two less. If the calculated amount of

nitric acid had formed, the pH would have dropped to 3.1. Thus several

anomalous results appeared in tests G-113 and G-114: the fixed nitrogen

concentration was much lower than calculated; the pH was also much lower; and

large concentrations of chloride appeared, the source of which was unknown.

Bates performed leach tests on the gasket material at 1500C in deionized water

for seven days and found no chloride release.22 He also performed leach tests

at 1500C for seven days on a stainless steel vessel using deionized water and

a pH 2 EN03 solution. In these tests he found 14.2 micrograms of C released

into the deionized water and 33.7 micrograms released into the KNO3 solution.

These results led us to suspect that the Cl in tests G-113 and G-114 came

from the vessels.

A brief study of the chemistry of steel making and consultation with

several steel chemists convinced us that chloride should not be present in

significant amounts in bulk steel, because of the high vapor pressure of

41

chlorides at the melting temperature of steel (see Section 3.14 for results of

stainless steel analysis). We then looked into the composition of the cutting

fluid used to fabricate the vessels and learned that Trimsol consists of an

emulsion of a chlorinated oil in water. The chlorine composition of the

concentrated emulsion sold by Master Chemical Company for water dilution by

users is about 10.5 wt%. Some of this material may have been buried in the

surface during machining and may have survived the cleaning process.

In view of these observations, it appears that the following processes

took place in vessels G-113 and G-114. Irradiation produced nitric acid in

the solution. The pH dropped, because there was no buffering. The acid

conditions accelerated the corrosion of the stainless steel and formed a

positively-charged Fe(III)-containing colloid. The corrosion released some

form of chlorine that had been buried in the surface of the stainless steel.

The chlorine was either initially present in the surface as chloride ions or

was reduced to chloride ions by the H and e radicals, forming H+ as well.aq

This led to more rapid corrosion. Some of the anions were attracted and

adsorbed by the colloid, which then coagulated and removed them from the

solution, forming the rust-colored precipitate. The absence of NO2 in the

solution resulted from the acid pH, as discussed in Section 3.11.

The behavior of tests G-265 and G-266 was more similar to the tests with

lower ratios of air to water than was the behavior of G-113 and G-114. The

predicted increase in fixed nitrogen concentration was 2.7 x 10 4 M, compared

to the observed values of 1.8 x 104 M and 7.5 x 10 5 M. The agreement is not

very close, but it is noteworthy that these two tests showed the only

measurable increases in fixed nitrogen in the second set of tests. The

differences are probably accounted for by the uncertainties discussed in

Section 3.9. The fact that all these values, both calculated and measured,

are below the bicarbonate concentration in equilibrated J-131 water explains

why the pH remained near neutral and no significant corrosion occurred.

3.14 ANOMALOUS TESTS

Three tests in the second set of experiments (out of a total of 66) also

showed anomalous results: G-208, G-247, and G-248. In all three cases, the

irradiation lasted for 91 days. At the end of the tests, the pH values were

found to be between 3.6 and 4.8; a flocculent, rust-colored precipitate was

42

observed the chloride levels were high (between 73 and 287 ppm); and other

anion concentrations were generally lower than usual. The fixed nitrogen had

been lost, rather than added to the solutions.

As mentioned in Section 2, other tests in this series produced solutions

that were yellow in color, while other parameters were normal. The explana-

tion of these results is probably the same as for the tests described in

Section 3.13. Apparently enough chlorine was accessible to produce the

effects even though a buffered solution, rather than deionized water, was

present. Analysis of the stainless steel of vessel 90, used in test G-247,

revealed 0.2 ppm Cl by weight in the bulk and a range of 0.5 to 300 ppm at the

surface, with an average of 5 ppm on the surface.22

43

44

4. CONCLUSIONS

It appears from this analysis that the observable radiation chemical

changes that would have been expected in those vessels of Bates et al. that

contained 16 ml of water at the outset of the experiments are the following:

1. Decomposition of a small fraction of the water to hydrogen and oxygen,

resulting in an observable pressure increase in the highest dose

experiments.

2. Fixation of a small amount of nitrogen from the air as nitrite and

nitrate ions in the solution. The increase in total fixed nitrogen

should have been measurable in the highest dose experiments.

3. Addition of an equivalent amount of hydrogen ions to the solution, which

would be well within the bicarbonate buffering capacity, and which should

therefore lower the pH only slightly from what it would be without

irradiation.

4. Variation in the nitrite-nitrate ratio, depending on the presence of

materials catalytic to the decomposition of hydrogen peroxide. A higher

ratio should be observed when catalysts are present.

5. Formation of Fe(III)-containing colloidal material from vessel wall

reactions.

It does not appear that measurable amounts of carboxylates (formate or

oxalate) should have been formed. The pH changes, though subtle, could

influence actinide solubility so that even the small changes caused by

irradiation could produce changes in the amounts of actinides in solution. It

appears from an examination of the Bates et al. data that these expectations

are largely borne out, except in a few tests that showed anomalously high

chloride levels. This chloride most likely came from residues from

chlorinated oil cutting fluid buried under the surface of the vessel walls.

This emphasizes the need to carefully specify the cutting fluids and cleaning

procedures to be used for waste packages.

45

46

5. APPLICATION TO THE REPOSITORY

The repository system is expected to differ in a number of ways from the

configuration used in the Bates et al. experiments. This is to be expected,

since the Bates et al. experiments were designed to represent saturated

conditions and to permit control of variables and convenience of operation and

analysis, while other engineering considerations apply to the repository

case. In applying the experimental results, one must keep these differences

in mind:

1. As noted in the Introduction, we expect that most of the nuclear waste

glass in the repository would remain dry during the first 300 to 1000

years-when significant gamma ray fluxes would be present. Under these

conditions, aqueous dissolution of the glass would not occur.

2. The repository would be a physically open system, because of the

permeability of the rock. Therefore, gases could escape from the

irradiated region, and the total gas pressure would remain at about 0.1

MPa. In addition, any liquid present would be free to migrate through the

rock, subject to the constraints of capillarity, permeability, and other

factors.

3. The ratio of gas to liquid volume after cooling to below the boiling point

could be considerably higher in the region near the containers in the

repository than in these experiments, particularly if there were only a

thin liquid film on the solid surfaces in contact with a relatively large

gas space in the annular region around each package.

4. The tuff would have a relatively larger effective surface area in the

repository in relation to the areas of metal and glass surfaces than it

did in these experiments.

5. The tuff surface in the repository would be essentially in contact with

the gas phase (perhaps through a thin liquid film) rather than being

submerged under bulk water.

6. The times of interest for the repository would be much longer than those

used in the experiments (100,000 years versus a few months), but the gamma

dose rates would decrease substantially with time (initially with a

30-year half-life) and with distance away from the packages (becoming

negligible in tens to hundreds of centimeters).

47

7. The temperatures in the repository would vary over space and time, rather

than being held at a fixed value.

8. The physical dimensions of the waste containers would be much larger than

those of the experimental vessels.

9. The repository design may incorporate hole liners made of carbon steel to

facilitate retrievability, as required in NRC regulations. This material

was not present in the experiments.

These differences and the uncertainties currently surrounding them make

the repository a more difficult case to analyze. Nevertheless, the main

features seen in the experiments should also be present in the repository:

1. Nitrogen in the gas phase would be fixed in the form of NO at the highest

temperatures, then N02 and N204 at lower temperatures, and finally as HN03

at temperatures near the boiling point of water. Some of this gas would

migrate to moist rock, dissolve, and undergo ion exchange with the

feldspars to produce nitrate and nitrite salts in the tuff. Some would

react with the metal surfaces to increase their oxidation rates. When the

surfaces had cooled so that a liquid film could exist on them, some of the

nitrogen oxide gases could dissolve in the film; at that time, the gamma

dose rates would be less than a few rad/hr for most of the packages. For

those that cooled to below the boiling point while the gamma dose rate was

still relatively high, the concentration of fixed nitrogen in the liquid

film could increase measurably, particularly if the ratio of gas to liquid

volume were large.

2. An equivalent amount of hydrogen ions would be produced in the liquid

film. If this amount did not exceed the buffering capacity of the

bicarbonate, the pH would not change significantly. If it did, the pH

would move in the acid direction, increasing the corrosion rates of the

metals.

3. The ratio of nitrite to nitrate would vary, depending on the surface upon

which the liquid film rested.

4. Carboxylates (formate and oxalate) should not form in liquid films in

contact with tuff or in those which had well-buffered pH due to dissolved

bicarbonate in concentrations in excess of those of radiation-produced

hydrogen ions. However, small amounts of carboxylic acids could be

expected to form in liquid films that reached low pH as a result of excess

48

nitric acid formation. If they formed, they could increase corrosion

rates by chelation.

5. The spatial variation of temperature coupled with the large physical size

of the packages could give rise to transport and concentration of

corrosive species by evaporation and condensation. However, the volume of

the resulting liquid is expected to be small.

6. A carbon steel hole liner, if present, could be expected to scavenge a

good share of the corrosive species.

7. Where liquid water containing dissolved species was irradiated,

decomposition to hydrogen and oxygen would occur. The rate would depend

on the radiation dose rate and would not be constrained by buildup of gas

pressure.

As the design parameters of the repository become more firmly fixed, it

will be possible to make estimates that are more quantitative of the radiation

chemical effects to be expected. In the meantime, the experiments of Bates et

al. have provided a number of significant insights.

49

50

REFERENCES

1. R. G. Baxter, Description of Defense Waste Processing Facility Reference

Waste Form and Canister, E. . duPont de Nemours and Co., Savannah River

Plant, Aiken, SC, DP-1606 Rev. 1 (1983).

2. L. R. Pederson and G. L. cVay, Influence of Gamma Irradiation on

Leaching of Simulated Nuclear Waste Glass: Temperature and Dose Rate

Dependence in Deaerated Water," J. Am. Ceram. Soc. 66, 863 (1983).

3. W. Primak and L. H. Fuchs, "Transportation of Matter and Radioactivity by

Ionized Air Corrosion," Physics Today 7, 15 (1954).

4. W. Primak and L. H. Fuchs, Nitrogen Fixation in a Nuclear Reactor,"

Nucleonics 13, 38 (1955).

5. J. Wright, L. K. Linacre, W. R. Marsh, and T. H. Bates, "Effect of

Radiation on Heterogeneous Systems of Air or Nitrogen and Water," in

Proc. 1st Int'l. Conf. Peaceful Uses Atomic Energy, Geneva, 1955 (United

Nations, Geneva, 1956), vol. 7, p. 560.

6. A. R. Jones, "Radiation-Induced Reactions in the N2-02-H20 System,"

Radiation Research 10, 655 (1959).

7. W. G. Burns, A. E. Hughes, J. A. C. Marples, R. S. Nelson, and

A. M. Stoneham, "Effects of Radiation on the Leach Rates of Vitrified

Radioactive Waste," J. Nucl. Mater. 107, 245 (1982).

8. W. M. Garrison, D. C. Morrison, J. G. Hamilton, A. A. Benson, and

M. Calvin, Reduction of Carbon Dioxide in Aqueous Solutions by Ionizing

Radiation," Science 114, 416 (1951).

9. N. Getoff, G. Scholes, and J. Weiss, Reduction of Carbon Dioxide in

Aqueous Solutions under the Influence of Radiations Tetrahedron Letters

-18, 17 (1960).

51

10. N. Getoff, Synthese organischer Stoffe aus ohlensiure in wsseriger

Lbsung unter Einwirkung von Co60 Gamma-Strahlung," Int'l. J. Appl.

Radiation and Isotopes 13, 205 (1962).

11. Aaron Barkatt, Alisa Barkatt, and W. Sousanpour, "Effects of Gamma

Radiation on the Leaching Kinetics of Various Nuclear Waste-form

Materials," Nature 300, 339 (1982).

12. Aaron Barkatt, Alisa Barkatt, and W. Sousanpour, "Gamma Radiolysis of

Aqueous Media and Its Effects on the Leaching Processes of Nuclear Waste

Disposal Materials," Nucl. Technol. 60, 218 (1983).

13. 0. Tamm, Experimentelle Studien uber die Verwitterung und Tonbildung von

Feldspaten," Chem. Erde 7, 420 (1930).

14. C. W. Correns and W. von Engelhardt, "Neue Untersuchungen uber die

Witterung des Kalifeldspates," Chem. Erde 12, 1 (1938).

15. C. W. Correns, "Die Chemische Verwitterung der Silikate," Naturw. 28, 369

(1940).

16. V. E. Nash and C. E. Marshall, "The Surface Reactions of Silicate

Minerals - Part II. Reactions of Feldspar Surfaces with Salt Solutions,"

Res. Bull. 614, Univ. of Missouri, Agric. Expt. Station, Columbia, MO

(1956).

17. R. M. Garrels and D. F. Howard, "Reactions of Feldspar and Mica with

Water at Low Temperature and Pressure," in Proc. of the Sixth Nat. Conf.

on Clays and Clay Minerals, Berkeley, CA, Aug. 19-23, 1957 (Pergamon

Press, New York, 1957), p. 68.

18. J. K. Bates and V. M. Oversby, "The Behavior of Actinide Containing Glasses

during Gamma Irradiation in a Saturated Tuff Environment," in Proc. of the

Materials Research Society Symposium on the Scientific Basis for Nuclear

Waste Management VIII, Boston, MA, Nov. 26-29, 1984, C. M. Jantzen et al.,

Eds. (Materials Research Society, Pittsburgh, PA, 1985), p. 257.

52

19. J. K. Bates, D. F. Fischer, and T. J. Gerding, The Reaction of Glass in a

Gamma Irradiated Saturated Tuff Environment. Part 1: SRL 165 Glass,

Argonne National Laboratory, Argonne, IL, ANL-85-62 (1985).

20. Environmental Protection Agency, Environmental Standards for the

Management and Disposal of Spent Nuclear Fuel, High-Level and Transuranic

Radioactive Wastesi Final Rule," in Code of Federal Regulations, Title 40,

Part 191, Federal Register Vol. 50, No. 182, p. 38066 (September 19, 1985).

21. Nuclear Regulatory Commission, Disposal of High-Level Radioactive Wastes

in Geologic Repositories," in Code of Federal Regulations, Title 10,

Part 60 (U.S. Government Printing Office, Washington, D.C., 1984).

22. J. K. Bates, Argonne National Laboratory, Argonne, IL, private

communication (1985).

23. R. A. Zielinski, Evaluation of Ash-flow Tuffs as Hosts for Radioactive

Waste: Criteria Based on Selective Leaching of Manganese Oxides, U.S.

Geological Survey, Denver, CO, Open File Report 83-480 (1983).

24. C. Willis, A. W. Boyd, and M. J. Young, wRadiolysis of Air and Nitrogen-

Oxygen Mixtures with Intense Electron Pulses: Determination of a

Mechanism by Comparison of Measured and Computed Yields," Can. J. Chem.

48, 1515 (1970).

25. C. Willis and A. W. Boyd, "Excitation in the Radiation Chemistry of

Inorganic Gases," Int'l. J. Radiat. Phys. Chem. 8, 71 (1976).

26. R. G. Macdonald and 0. A. Miller, "Low Dose-Rate Radiolysis of

Nitrogen: Yield of Nitrogen Atoms, N( S) and N(2D,2P)," Radiat. Phys.

Chem. 26, 63 (1985).

27. 0. Tokunaga and N. Suzuki, "Radiation Chemical Reactions in NOx and SO2

Removals from Flue Gas," Radiat. Phys. Chem. 24, 145 (1984).

53

28. F. Busi, M. D'Angelantonio, Q. G. Mulazzani, V. Raffaelli, and

0. Tubertini, "Radiation Treatment of Combustion Gases: Formulation and

Test of a Reaction Model,' Radiat. Phys. Chem. 25, 47 (1985).

29. S. Gordon, K. H. Schmidt, and J. R. Honekamp, "An Analysis of the

Hydrogen Bubble Concerns in the Three-Mile Island Unit-2 Reactor Vessel,"

Radiat. Phys. Chem. 21, 247 (1983).

30. W. G. Burns and P. B. Moore, "Water Radiolysis and Its Effect upon

In-Reactor Zircaloy Corrosion," Rad. Effects 30, 233 (1976).

31. A. W. Boyd, M. B. Carver, and R. S. Dixon, "Computed and Experimental

Product Concentrations in the Radiolysis of Water," Radiat. Phys. Chem.

15, 177 (1980).

32. H. Christensen and E. Bjergbakke, "Radiolysis of Ground Water from Spent

Fuel," Svensk Karnbransleforsorjning AB/Avdelning KBS, Stockholm, Sweden,

SKBF KBS Report No. 82-18 (1982).

33. S. A. Simonson and W. L. Kuhn, "Predicting Amounts of Radiolytically

Produced Species in Brine Solutions," in Proc. of the Mater. Research

Soc. Symp. on the Sci. Basis for Nucl. Waste Management, Boston, Mass.,

Nov. 14-17, 1983, G. L. McVay, Ed. (Elsevier, New York, 1984), p. 781.

34. S. L. Nicolosi, "A Generalized Model for the Analysis of Groundwater

Radiolysis," in Proc. of the Mater. Res. Soc. Symp. on the Sci. Basis for

Nucl. Waste Management VIII, C. M. Jantzen et al., Eds. (Materials

Research Society, Pittsburgh, PA, 1985), p. 631.

35. T. Eriksen and A. Jacobsson, "Radiation Effects on the Chemical

Environment in a Radioactive Waste Repository," Svensk

Karnbransleforsorjning AB/Avdelning BS, Stockholm, Sweden, SKBF KBS

Report No. 83-27 (1983).

54

36. W. G. Burns, W. R. Marsh, and W. S. Walters, The Gamma Irradiation-

Enhanced Corrosion of Stainless and Mild Steels by Water in the Presence

of Air, Argon, and Hydrogen," Radiat. Phys. Chem. 21, 259 (1983).

37. J. W. T. Spinks and R. J. Woods, An Introduction to Radiation Chemistry

(Wiley, New York, 1976), 2nd ed.

38. C. J. ochanadel and J. A. Ghormley, Effect of Temperature on the

Decomposition of Water by Gamma Rays," Radiat. Res. 16, 653 (1962).

39. G. G. Jayson, B. J. Parsons, and A. J. Swallow, "Some Simple, -Highly

Reactive, Inorganic Chlorine Derivatives in Aqueous Solution," J. Chem.

Soc. Faraday Trans. I 69, 1597 (1973).

40. D. Behar, G. Czapski, and I. Duchovny, Carbonate Radical in Flash

Photolysis and Pulse Radiolysis of Aqueous Carbonate SolutionsW J. Phys.

Chem. 74, 2206 (1970).

41. G. E. Adams, J. W. Boag, and B. D. Michael, Transient Species Produced

in Irradiated Water and Aqueous Solutions Containing Oxygen," Proc. Roy.

Soc. London, Ser. A 289, 321 (1965).

42. E. Hayon and J. J. McGarvey, Flash Photolysis in the Vacuum Ultraviolet

Region of S042 , cO3 , and OH Ions in Aqueous Solutions," J. Phys. Chem.

n, 1472 (1967).

43. T. E. Eriksen, J. Lind, and G. erenyi, On the Acid-Base Equilibrium of

the Carbonate Radical,* Radiat. Phys. Chem. 26, 197 (1985).

44. K. V. Baczko, Ed., Gmelin andbuch der Anorganischen Chemie, System-

Nummer 14, Part C, Number 3, "Kohlenstoff" (Verlag Chemie-GMBH,

Weinheim/Bergstrasse, Germany, 1973), p. 123.

45. F. A. Cotton and G. Wilkinson, Advanced Inorganic Chemistry (Wiley, New

York, 1980).

55

46. W. J. Gray and S. A. Simonson, "Gamma and Alpha Radiolysis of Salt

Brines," in Proc. of the Materials Research Society Symp. on the Sc.

Basis for Nucl. Waste Management VIII, Boston, Mass., Nov. 26-29, 1984,

C. M. Jantzen et al., Eds. (Materials Research Society, Pittsburgh, PA,

1985), p. 623.

47. A. Fojtik, G. Czapski, and A. Henglein, "Pulse Radiolytic Investigation

of the Carboxyl Radical in Aqueous Solution," J. Phys. Chem. 74, 3204

(1970).

48. J. P. Keene, Y. Raef, and A. J. Swallow, "Pulse Radiolysis Studies of

Carboxyl and Related Radicals," in Pulse Radiolysis, Proc. of the Int'l.

Symp., Manchester, England, April, 1965 (Academic, New York, 1965), p. 99.

49. I. G. Draganic and 0. Gal, Radiation Chemistry of Oxalic Acid and

Oxalates,l Rad. Res. Rev. 3, 167 (1971).

50. H. Stephen and T. Stephen, Solubilities of Inorganic and Organic

Compounds, Vol. 1, Binary Systems, Part 1 (Macmillan, New York, 1963),

p. 251.

51. T. Hasselstrom and M. C. Henry, "New Synthesis of Oxalic Acid," Science

123, 1038 (1956).

52. M. C. Juhas, R. D. McCright, and R. E. Garrison, Behavior of Stressed and

Unstressed 304L Specimens in Tuff Repository Environmental Conditions,

Lawrence Livermore National Laboratory, UCRL-91804 (1985) (presented at

Corrosion '85, National Association of Corrosion Engineers Annual

Meeting, Boston, MA, March 25-29, 1985).

53. R. S. Glass, R. A. Van Konynenburg, and G. E. Overturf, Corrosion

Processes of Austenitic Stainless Steels and Copper-Based Materials in

Gamma-Irradiated Aqueous Environments," Lawrence Livermore National

Laboratory, UCRL-92941 (1986) (presented at Corrosion '86, National

Association of Corrosion Engineers Annual Meeting, Houston, Texas, March

17-21, 1986).

56

54. W. C. Schumb, C. N. Satterfield, and R. L. Wentworth, Hydrogen Peroxide

(Reinhold, New York, 1955).

55. J. K. Linacre and W. R. Marsh, "The Radiation Chemistry of Heterogeneous

and Homogeneous Nitrogen and Water Systems," Chemistry Division, AERE

Harwell, England (1981), AERE-R10027.

56. J. A. Roth and D. E. Sullivan, "Solubility of Ozone in Water," Ind. Eng.

Chem. Fundam. 20, 137 (1981).

57. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions

(Pergamon, New York, 1966).

58. R. A. Van onynenburg, Lawrence Livermore National Laboratory,

unpublished data discussed in internal memorandum entitled Water

Chemistry Results from Corrosion Experiment No. 11 (General Corrosion of

Stainless Steels in Hot, Unirradiated J-13 Water)," October 22, 1985.

59. W. L. Jolly, The Inorganic Chemistry of Nitrogen (W. A. Benjamin, New

York, 1964).

60. K. L. Nash, S. Fried, A. M. Friedman, N. Susak, P. Rickert, and

J. C. Sullivan, Radiation Effects in Solution and on the Solid/Liquid

Interfaced Nucl. Technol. 60, 257 (1983).

61. R. C. Weast and M. J. Astle, Eds., CRC Handbook of Chemistry and Physics

(CRC Press, Boca Raton, FL, 1982), 63rd ed.

62. J. . Keenan, F. G. Keyes, P. G. Hill, and J. G. Moore, Steam Tables

(Wiley, New York, 1978).

63. J. A. Dean, Ed., Lange's Handbook of Chemistry (McGraw-Hill, New York,

1979), 12th ed.

64. A. M. Amorosi and J. R. McDermet, The Calculation of the Distribution of

Carbon Dioxide Between Water and Steam," ASTM Proc. 39, 1204 (1939).

57

65. P. Cohen, Water Coolant Technology of Power Reactors (American Nuclear

Society, La Grange Park, IL, 1969).

58

Technical Information Department Lawrence Livermore National LaboratoryUniversity of California Livermore, California 94550